Commercial training data from specialist providers. Listings show what each source advertises; licensing is confirmed with the provider before any agreement.
This dataset contains over 250,000 hours of podcast dialogues recordings, collected from contributors around the world.
A large-scale multimedia dataset containing 30K+ hours of official government communications recordings captured globally, designed for speech, video, and AI…
This dataset contains over 500,000 hours of call center audio recordings, collected from contributors around the world.
This dataset contains over 500,000 hours of multilingual conversations recordings, collected from contributors around the world.
This dataset contains over 50,000 hours of educational lectures recordings, collected from contributors around the world.
A large-scale multimedia dataset containing 5K+ hours of multi-sport audio commentary recorded across diverse sporting events and broadcast environments.
1. Specifications Format : 8kHz, 8bit, u-law/a-law pcm, mono channel; Environment : quiet indoor environment, without echo; Recording content : No preset linguistic…
Specifications: ‣ Volume: 100,000+ hours ‣ Format: 8kHz, mono channel (WAV audio + JSON transcript) ‣ Environment: Studio, podcast-quality recording, podcast audio data…
This dataset offers a large-scale multilingual speech corpus covering 54 global languages, including Arabic, Hindi, English (US/UK/India), Mandarin, Spanish, French, and…
Specifications: ‣ Volume: 100,000+ hours ‣ Format: 8kHz, mono channel (WAV audio + JSON transcript) ‣ Environment: Studio, podcast-quality recording ‣ Recording content:…
Specifications: ‣ Volume: 100,000+ hours ‣ Format: 8kHz, mono channel (WAV audio + JSON transcript) ‣ Environment: Studio, podcast-quality recording ‣ Recording content:…
Specifications: ‣ Volume: 100,000+ hours ‣ Format: 8kHz, mono channel (WAV audio + JSON transcript) ‣ Environment: Studio, podcast-quality recording ‣ Recording content:…
Clean-licensed Indonesian conversational speech, without the gray zones.
Specifications: ‣ Volume: 100,000+ hours ‣ Format: 8kHz, mono channel (WAV audio + JSON transcript) ‣ Environment: Studio, podcast-quality recording, call center audio ‣…
A large-scale multimedia dataset containing 100,000 hours of scripted dialogue from long-form television series and soap operas, performed by professional actors in…
A large-scale multimedia dataset containing 10K hours of local news interviews recorded across diverse regions, designed for speech, video, and AI experimentation.
34.5K hours of OTS owned & operated, rights-cleared conversational audio across hundreds of speakers and multiple US regions.
WiserBrand offers premium audio data sourced exclusively from real-world, high-emotion customer service conversations.
High-quality conversational audio dataset collected through a global network of 3M+ verified data collectors across 150+ countries.
This dataset contains thousands of authentic audio recordings of customer calls to service teams across key U.S.
TranscriptDB is the full YouTube transcript firehose, built for systematic hedge funds with in-house NLP capacity and AI labs training foundation models.
This dataset includes a large-scale collection of podcast audio data across multiple languages, domains, and speaker demographics.
The provider provides licensed, consent-verified speech recordings from native speakers for AI/ML model training.
Scope: Synthetic customer–agent interactions, available in dual-channel (per speaker) or single-channel mixes.